RMS Weighs in on BitKeeper Debacle 1137
mshiltonj writes "You know its what we've all been waiting for: RMS weighs in on the BitKeeper debacle. An excerpt: "I want to thank Larry McVoy. He recently eliminated a major weakness of the free software community, by announcing the end of his campaign to entice free software projects to use and promote his non-free software. Soon, Linux development will no longer use this program, and no longer spread the message that non-free software is a good thing if it's convenient."
Re:yeeeeeeeeha!!! (Score:2, Informative)
Re:Umm... (Score:5, Informative)
Re:he's being quite modest about it (Score:1, Informative)
Re:Strange.. (Score:5, Informative)
Uh-huh, right......
BitKeeper is not "open source." Nobody ever got the source outside of Larry McVoy's company. BitKeeper is proprietary software that you normally have to pay money to use. McVoy allowed "free" use for "free" software projects and Linus chose to use it for managing his end of Linux kernel development.
After Andrew Tridgell showed how you could connect to a BitKeeper repository using netcat to see what the "protocol" does, Mr. McVoy said no more "free" BitKeeper for you and went home.
No Open Source or Free Software projects were harmed in all of this, except that now Linus is going to develop his own tool for managing the kernel code instead of using something that's already available, because apparently, he's tried them all and decided that none really work for him.
Re:I hate RMS (Score:3, Informative)
Re:Why (Score:1, Informative)
RMS and games (Score:2, Informative)
His answer was that in those cases they can have it closed source for a few months (3?) and then release it open source.
I guess a trend that might be possible today would be to have open source engines and pay-for content.
Re:the cost of innovation (Score:3, Informative)
Wrong, wrong, and more wrong. You don't understand what you quoted. Tridge didn't write a replacement for BitKeeper. He wrote a tool that allows you interoperate with BitKeeper - to get the source code out of BK without using BK.
Re:Usually incisive, RMS emphasizes the wrong poin (Score:5, Informative)
No, he can't.
From the FAQ [gnu.org]
Linus can redistribute code he has written under another license, but he cannot revoke the rights he has already provided. He can also make it so future releases are under a more restrictive license, but someone would just end up forking the last GPLed version.
A good example of this is XFree86. Version 4.4 was released under a more restrictive license that the community did not like. Next thing you know, the last 4.4 prerelease under the old license was forked as X.org.
Re:he's being quite modest about it (Score:3, Informative)
You're going to have a very long wait. Love it or loathe it,
Re:Usually incisive, RMS emphasizes the wrong poin (Score:4, Informative)
Read the licence. There is no provision for retroactive modification or revocation of the license. It is an outright grant of permission. It cannot be be withdrawn, because there is no basis in the license for doing so.
HOWEVER (and this is the point you seem to be making) an author may license his software to different people under different licenses. If I license my program to you under the GPL and to Apple under a Microsoft-style EULA, you will still have all the GPL rights granted to you. In fact, Apple could have gotten the software from you under the GPL (which is the only license YOU can distribute my program under, since you license it and do not own it). Apple could then distribute the software under the GPL (and the people who got it from Apple...ad infinitum), but Apple would be bound by the GPL with regard to modifications that they distribute. Apple doesn't like that, so they come to me with money and a request for a different license.
But no matter what happens between me and Apple, between me and you was the GPL. You still have the GPLed copy of my software, and if I go capitalistic nuts tomorrow and begin demanding $1000/day before I'll distribute any more copies of my program, you would still be able to use, copy, modify, and redistribute the copy of my program that I gave to you.
I realize at this point that I am arguing by repeated assertion, so I encourage you again to go read the license yourself. Note that there is NO basis for revocation or modification of the license. It is a contract, and American contract law doesn't permit unilateral modification of contracts. (If it did, I might modify my mortgage contract.)
Re:A question for RMS (Score:4, Informative)
http://c2.com/cgi/wiki?TecoEditor
Re:he's being quite modest about it (Score:3, Informative)
While your first statement on the status of free java is true, the second one isn't quite. The public libraries are pretty well documented.
The #1 problem OOo has with Free Java is that the Sun hackers don't use the public libraries. They make use of Sun-specific internal libraries which are not publicly documented at all, and are not supposed to be used by applications. These libraries aren't part of Java.
To give a practical example, there are classes like com.sun.net.ssl.internal.ssl.Provider , (code using which I've seen code floating around). This class is not available on the Apple or IBM java runtimes. (Both of which are Sun-approved as 'Java')
So it's not as much that they're using Java as the fact that they're not using Java as it's intended but rather Java coded to only work with the Sun JRE specifically.
Anyway.. So far Red Hat has been working on compiling the parts of OOo that do work (or can be made to work) with GCJ for shipping with their distro. I suspect Debian and so on will do something similar. So in that sense, it's already forked.